Research map / Architectures and efficiency / CNN design
GPCA: A Probabilistic Framework for Gaussian Process Embedded Channel Attention
Jiyang Xie, Zhanyu Ma, Dongliang Chang, Guoqiang Zhang, Jun Hai Guo
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add Gaussian process prior to channel attention (ablation) | GPCA (ResNet50) · CIFAR-100 | top-1 accuracy: 72.36 -> 73.93 | Improved |
| add Gaussian process prior to channel attention (ablation) | GPCA (ResNet18) · ImageNet-32x32 | top-1 accuracy: 47.37 -> 49.58 | Improved |
| add Gaussian process prior to channel attention (ablation) | GPCA (VGG16) · miniImageNet | accuracy: 81.30 -> 82.78 | Improved |
| optimize kernel parameters instead of fixing them (ablation) | GPCA (ResNet34) · miniImageNet | accuracy: 79.92 -> 80.14 | Improved |
| optimize kernel parameters instead of fixing them (ablation) | GPCA (VGG16) · CIFAR-10 | top-1 accuracy: 94.50 -> 94.51 | No clear effect |
| optimize kernel parameters instead of fixing them (ablation) | GPCA (ResNet18) · ImageNet-32x32 | top-1 accuracy: 49.34 -> 49.58 | Improved |
| add GPCA module to base network | ResNet50 · CIFAR-100 | top-1 accuracy: 71.88 -> 73.93 | Improved |
| add GPCA module to base network | ResNet34 · miniImageNet | accuracy: 78.08 -> 80.14 | Improved |
| add GPCA module to base network | ResNet18 · ImageNet | top-1 accuracy: 70.40 -> 71.61 | Improved |
| add GPCA module to base network | ResNet50 · ImageNet | top-1 accuracy: 75.20 -> 77.57 | Improved |
| replace SENet with GPCA | ResNet18 · ImageNet-32x32 | top-1 accuracy: 47.89 -> 49.58 | Improved |
| add GPCA to ADL for localization | ADL (VGG16) · CUB-200-2011 WSOL | top-1 localization accuracy: 45.30 -> 54.97 | Improved |
Closest papers by meaning
- Gaussian Context Transformer (CVPR 2021 · similarity 0.87)
- MCA: Moment Channel Attention Networks (AAAI 2024 · similarity 0.85)
- ECA-Net: Efficient Channel Attention for Deep Convolutional Neural Networks (CVPR 2020 · similarity 0.84)
- Attention-based PCA (NeurIPS 2026 · similarity 0.83)
- Degrees of Freedom for Linear Attention: Distilling Softmax Attention with Optimal Feature Efficiency (NeurIPS 2025 · similarity 0.83)
- Gated Linear Attention Transformers with Hardware-Efficient Training (ICML 2024 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
